Brighton and Hove Secondary School Admissions Proposal - Week 3 Analysis
Author
Professor Adam Dennett FRGS FAcSS, Professor of Urban Analytics, Bartlett Centre for Advanced Spatial Analysis, University College London - a.dennett@ucl.ac.uk - @adam_dennett
Reading layer `BrightonLSOA_Clean' from data source
`D:\BH_Schools_2\data\BrightonLSOA_Clean.geojson' using driver `GeoJSON'
Simple feature collection with 165 features and 27 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: -0.2450499 ymin: 50.79908 xmax: -0.0160027 ymax: 50.89237
Geodetic CRS: WGS 84
Reading layer `lsoa_btn_pw_cent' from data source
`D:\BH_Schools_2\data\lsoa_btn_pw_cent.geojson' using driver `GeoJSON'
Simple feature collection with 165 features and 1 field
Geometry type: POINT
Dimension: XY
Bounding box: xmin: 524624.9 ymin: 102198.2 xmax: 538674.1 ymax: 109563.7
Projected CRS: OSGB36 / British National Grid
Reading layer `LSOA_NOMIS_plus_projections' from data source
`D:\BH_Schools_2\data\LSOA_NOMIS_plus_projections.geojson'
using driver `GeoJSON'
replacing null geometries with empty geometries
Simple feature collection with 3465 features and 5 fields (with 3465 geometries empty)
Geometry type: GEOMETRYCOLLECTION
Dimension: XY
Bounding box: xmin: NA ymin: NA xmax: NA ymax: NA
Geodetic CRS: WGS 84
Reading layer `BrightonSecondaryCatchments' from data source
`D:\BH_Schools_2\data\BrightonSecondaryCatchments.geojson'
using driver `GeoJSON'
Simple feature collection with 6 features and 2 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 523887.8 ymin: 100896.4 xmax: 539613.4 ymax: 110190.1
Projected CRS: OSGB36 / British National Grid
Reading layer `optionA' from data source `D:\BH_Schools_2\data\optionA.geojson' using driver `GeoJSON'
Simple feature collection with 6 features and 2 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 523889.3 ymin: 100892.8 xmax: 540295.5 ymax: 110191.6
Projected CRS: OSGB36 / British National Grid
Reading layer `optionB' from data source `D:\BH_Schools_2\data\optionB.geojson' using driver `GeoJSON'
Simple feature collection with 4 features and 2 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 523890.5 ymin: 100894.1 xmax: 539613.1 ymax: 110186.1
Projected CRS: OSGB36 / British National Grid
[1] "21.0.4"
Reading layer `oa_brighton' from data source
`D:\BH_Schools_2\data\oa_brighton.geojson' using driver `GeoJSON'
Simple feature collection with 936 features and 10 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 523598.1 ymin: 101780.1 xmax: 539862.4 ymax: 111919.1
Projected CRS: OSGB36 / British National Grid
I’m going to begin this week with a little experiment. Imagine, if you can, a Brighton where all the Secondary Schools are exactly the same. They are all the same size. They are all of the same quality - no more “Outstanding” or “Requires Improvement”. Just the same. All in identikit buildings, with the same allocation of green space. Same teachers. Same school dinners. They are identical in every way. The only difference between them is where they are located in the City. As it happens, they are all located exactly where they are today. Varndean is where Varndean is. Longhill is where Longhill is. Hove Park is where Hove Park is, this is the only thing to differentiate them.
Now imagine that all of the children in the city live where they currently live. But other than that, they are also all identical. Same size families, same levels of income, same religion, etc.
In this slightly weird version of Brighton - an equal Brighton (some might say a “Brightopia”) - which schools would the children in the city choose to attend if the ONLY thing that gave them preference, was how long it took them to get to school either by walking or taking the bus?
If we knew the answer to this question, we might then start to understand what role location and distance plays in the choices people make. And if we know this, we can then compare this “Brightopian” reality with the situation we have in real Brighton today.
Why is this useful? Well, if we can disentangle the spatial effects from all of the other effects (how ‘good’ the schools are, religious preference, what the lunches are like, which catchment they are in etc.) we can start to understand what impact things like school catchment areas or perceived quality, are having on the system. Is it the case that people want to go to this school because it is “good” or is it just because it is close to where they live?
A Model of Brightopia
Figure 1 - Brighton and Hove Secondary Schools and 11 year olds in 2024.
The map above is a real map of Brighton and Hove. The orange circles represent numbers of 11 year olds in the city in 2024 and where they live - these numbers are taken from the ONS sub-national populations at the city level and distributied down into neighbourhoods according to where the 2021 Census records where they live. This is a pretty accurate representation of where the children who were going to attend secondary shcool in 2024 live.
The black dots are the locations of the Secondary Schools in the city and the black boundaries are the current catchment areas surrounding these schools, for context.
You can see in the background for this map, there is also a road network. And if we know roughly how fast people walk and if we have information on the bus network in the city and the bus timetables associated with each bus-stop in the city (to give us an idea of how fast the buses travel), we can generate some pretty good estimates of how long it would take children in the orange circles to get to the black dots.
We can use this information to create a ‘model’ (a simplified version of reality) of Brighton - our “Brightopia” model and we can use this model to help us answer exactly the questions we asked above:
If the only thing that mattered was how long it took to get to school, which schools would the children in the city choose to attend?
Running the Brightopia model
In my day job, I build these kinds of models all the time. They are called ‘spatial interaction models’ and they are used to understand how people and commodities move around in cities. They are used to understand how people get to work, how they get to the shops, how they get to school, where they spend their money and how they make choices about where they go. Supermarket chains use them to predict store revenues, transport companies use them to predict passenger numbers, and urban planners use them to understand how cities work.
Spatial interaction models are sometimes simply referred to as ‘gravity models’. This is because the number of people moving between two places is proportional to the size of the two places and inversely proportional to the distance between them - a concept exactly the same as Sir Isaac Newton’s formula for gravity - the bigger the two objects and the closer they are, the stronger the force between them.
If you would like to understand a little more about the theory and the mathematics behind these models - I wrote an explainer some years ago here - https://rpubs.com/adam_dennett/376877
The model I have decided to run is called a “Production” or “Origin Constrained” model - what this means is that the number of children who can go to each school are constrained to the numbers we know are living in different neighbourhoods in the city. The numbers actually arriving at schools reflect demand based only on geography.
Which schools students decide to go to is then only determined by the time it takes to get to school and nothing else. This is a simplification of course - in reality, there are many other factors that determine where children go to school - but it is through this simplification that we uncover the importance of space.
I have gone a step further and constrained these numbers to the total PAN of 2560 11-year old places in the council’s published material so what we will end up with are estimates that we can compare with actual PAN numbers to see how close we get.
What does the Brightopia Model tell us?
If you don’t want to wade through my opinions, you may wish to jump straight to Some New Analysis and find out what this map is all about - otherwise, read on below.
Review to Date
Since the announcement of the ‘engagement exercise’ by the council, and frustrated by the evidence made available, I have made it my mission to try and fill the information vacuum with data and evidence collated, analysed and presented from a variety of different sources.
In the last couple of weeks, I have produced two pieces of analysis (here and here) which have helped zoom in where the real issues in the system are. These can be summarised as:
The spatial distribution of schools and places available in those schools, relative to where pupils actually live. Some (well, one as we’ll see below) of the schools in Brighton are simply in the wrong place (piece 1).
Illogical reductions in Published Admission Numbers for some schools and the maintenance of other, relative to current and future demand, driven largely by a desire to artificially prop up a school of the wrong size in the wrong location (piece 2).
Unfortunately, the effort taken in digging through the data and to unearly these issues means that I’ve not yet been able to address some of the other crucial parts of the story. In particular:
The fact that the educational disadvantage partially caused by the spatial arrangement of current schools, is most keenly felt by those in the city who unfairly suffer from other forms of social and economic disadvantage - in particular residents of Whitehawk whose access to secondary education is poor. The issues raised by Class Divide around the closure of the COMART school in East Brighton in 2005 and their subsequent lack of accessibility to ANY secondary education in the city, let alone high-quality education, is very real.
But in not being fully honest about all dimensions of problem - or possibly just ignorant of them - and in trying to portray the whole of the Brighton schools system as a problem, and a narrative of the ‘middle classes’ of north Brighton closing ranks to thwart a just cause (rather than a fundamentally flawed process), the whole community has been let down by the Council. I am now even more strongly of the opinion that the ‘solutions’ on the table are nothing of the sort and will only serve to do more harm than good.
My Current Fears - I hope unfounded
At the beginning of the exercise, I had real hope that the lack of information coming from the council was perhaps a mistake - an oversight due to an inexperienced or over-worked team. Unfortunately, as the weeks have gone by, and as I have listened to members of the council during the ‘engagement’ Q&A sessions, the council are giving the impression that the lack of information is intentional. By not engaging with the requests for information and the clearly articulated fears of the community - and at worst, being dismissive of them - it feels like they have already made up their minds. It feels like an engagement in name only. I hope this is not the case.
Having reviewed all of the data I have gathered, it has become clear that only one option on the table - Option B (or something which looks a lot like it) - will result in the mass transition of students from the centre of the City, to Longhill in the East. A school which while once an appropriate size for a catchment stretching out towards Peacehaven, thanks to long-term demographic shifts (an ageing population and decline in families) no longer serves its original constituency. In recent years it’s cohorts of students have been swelled by children bussed out from Whitehawk in the East of the City - a demographic sticking plaster made easier to apply thanks to a population with less financial and social capital to resist.
To its credit, the Council - if not quite able to own up to its failures in closing COMART in 2005 - at least is now listening to that part of the city and has enacted a new policy to allow the financially worst off students - those on Free School Meals - to choose to study elsewhere in the city, should they wish. However, it’s clear that Option B is also the only option that stands to mitigate the inevitable effects of this policy, which will lead to still more students deserting Longhill. A policy that, if the council do rush through with a consultation immediately after the engagement period ends, won’t have had time to come to fruition or the initial results to be reviewed. It can be the only explanation for the haste with which council is proceeding.
I fear that in all of this, if the council pushes forward against the increasing weight of evidence, the people they are supposedly championing - those particularly in Whitehawk - will actually lose out in the long term. This is because, as I will show in some new analysis below - Longhill is poorly located and will continue to be poorly located, relative to where pupils live in the city. The council may try to argue that a new housing estate in Ovingdean will generate the demand required to keep Longhill afloat in the medium term, but as the numbers show below, unless this estate is the size of something like Whitehawk (it is far from that) it will be a drop in the ocean and barely move the dial. And the distraction of making a bad catchment system even worse, means that a real, sustainable, long-term, solution will be kicked down the road in favour of a poorly thought through short term solution, which is unlikely to work anyway.
My Hopes
On the plus side, it is not too late for the council to take notice of the hundreds of good ideas and suggestions that are now emerging from individuals and self-organised groups in the city. I hope that they really do listen.
I hope that some of the evidence I have provided to fill the vacuum has helped some people understand what the underlying problem is a little better. And that armed with this information, they have felt emboldened to pose good questions to the council and make their voice heard.
I hope that the council takes the city up on its offer of help, whether through a Citizens Assembly, through the offers of help from experts who live in the city and care about it deeply.
In my last piece I offered the council a viable option which would allow them to slow down and take an evidence-based approach. Myself and other members of the newly formed Parent Support Group intend to present the council with a worked through alternative to their current plans - and “Option D” if you like which will offer both a short-term and longer term sustainable alternatives to those that are currently on the table.
I hope that our proposals do buy the time the Council needs to do this properly. In my last piece I called for a systems thinking approach and I think this is the only way a lasting and sustainable solution can be reached. This will need to be carried out with a full range of view gathered from a Citizens Assembly - and idea I have seen recently promoted by Class Divide and one which was put to the council the last time they tried to re-draw the catchment boundaries, but one which clearly has not been enacted yet. And the council need to engage with a diverse and full range of expert views. It is great that the view of Prof Gorard and Dr Greaves were heard, and their educational perspective is important, but as I have shown this is as much a problem of geography and community, of flows and interactions, as it is of education. And there are hundreds of other voices with equally important perspectives that need to be heard.
Some New Analysis
“A Stain on our City” - really?
The phrase “A Stain on our City” has been in the back of my mind ever since it was used in reference to one of the opening slides that Jacob Taylor produced during the face-to-face consultation meeting on Tuesday 8/10/2024. It is highly a highly charged statement used as an oratorical device in making his case. Somebody pulled him up on the emotive language in the audience. Clearly taking this onboard he then went on to use it again in the People, Overview and Scrutiny Committee Meeting - it can be seen here near the opening of his speech - https://aisapps.mediasite.com/AuditelScheduler/Player/Index/?id=c1841cfa-af61-45cf-85b5-bb8d274e6b75&presID=32ca1cfe29e14291be27a478fdbf76d01d
As a reminder, this is the graph that he was referring to when he made these comments:
A Stain On Our City?
The ‘stain’ in this context is in reference to the 24% of Brighton & Hove’s disadvantaged pupil population getting grade 5 or above in Maths and English. Being below the national figure of 25.4 for England constituted a “stain on our city” in JT’s mind. Strong stuff. Especially considering Brighton seems to be doing better than the rest of the South East region.
But is it really a “stain”, or is there more going on?
Now, of course, comparing these figures with non-disadvantaged pupils, there is clearly a big attainment gap and narrowing this gap must be (and I am sure is) a priority for every local authority in the country. It’s not right that those from disadvantaged backgrounds do so much worse than their more advantaged peers.
But the argument Jacob Taylor was employing cast the whole of Brighton as a failure, relative to England. But is it Brighton really doing as badly as cast?
Two questions arise from this:
Is 24% as bad as is being portrayed if we examine some wider context?
Is the gap between 24% and 50% worse or better than anywhere else in the country?
In answering both of these questions, we can develop a clearer context for Brighton - which is really important as it underpins a lot of the arguments being made by the Council.
Figure 1 - Percentage of Students achieving Grade 5+ English and Maths, for disadvantaged (those in receipt of Free School Meals - left) and non-disadvantaged (right) students, 2022-23.
There are a few features of note in this graph.
Firstly, there are 33 ‘Regions’ in this version of England. Brighton is within, and contributing to, the value of 21.4% for Surrey, East and West Sussex. Brighton, at 24%, is not only doing better than its wider region, but because it is not excluded from the regional figure, actually pulling its average up.
Secondly, if Brighton were a region on this scale (ranked according to FSM achievement), it would sit between West Yorkshire (23.8% - ranked 10 of 33) and Berkshire, Buckinghamshire and Oxfordshire (24.5% - ranked 9 of 33). In other words, it would fall in the top 1/3 in England.
Thirdly, you may have noticed something interesting at the top of the graph - the top 5-regions that perform considerably better on the attainment of their pupils in receipt of Free School Meals, are all parts of London.
Indeed London performs a full 10% better than the next ranked region and does a fantastic job of pulling the average up for the rest of the country. Which takes me back to the question I asked in my first piece of work on this topic - why is London doing much better than Brighton and almost everywhere else in the Country? A question I can’t answer here, but one incredibly relevant to this whole issue.
Forthly, without London, the average 5-9 GCSE rate for disadvantaged pupils for the rest of the country reduces to 21%. Against which Brighton is performing much better than average, and if it were one of these regions it would be, in the top 5.
Finally (and it might be easier to read across the table below to see this), the FSM / Non-FSM attainment gap in Brighton, according to the Council’s graph at the top of this piece, is 26%. As you can see in the table below - if Brighton were a region, this would rank 12th on this scale out of 33 - just outside the top 1/3.
Table 1 - Attainment Gap at Age 16, FSM Eligible and Non-Eligible Students, England 33 Regions. Ranked by Smallest Gap
Hopefully what I have shown in this short exercise, is that context is everything.
Compared the the national average for England, Brighton’s attainment at GCSE for pupils on Free School Meals is lower, however, that national average is being pulled up considerably by London, which is performing between 6.5% and 15.1% better than the best-performing non-London region.
If Brighton were one of these 33 regions, it woud still be achieving better results for pupils on FSM than 2/3 of other places in England. If we take London out of the equation, it would appear in the top 5.
Rather than being ‘a stain on our city’, it should be regared as a badge of pride that Brighon and Hove is able to achieve such good results for pupils in receipt of Free School Meals, relative to the rest of England.
If we look at the attainment gap between those who are on Free School Meals and those who aren’t as well, Brighton still performs far better than average, even with London in the equation. If Brighton were one of the regions above, it would be just outside the top 1/3.
The characterisation of Brighton and Hove’s ability to help it’s more deprived students as a ‘stain on the city’ is demonstrably unfair. And this kind of ‘cherry picking’ of statistics (where indeed statistics of any kind have been provided) to suit a particular line of argument has, sadly, characterised this whole ‘engagement’ process.
However, this is even more to it than this. That 24% potentially masks some particularly acute problems in parts of the City such as East Brighton - listening to the Class Divide Podcast, this could be as high as 32% in this part of the city. So ‘stain’ rhetoric when tarring the whole city plays a slightly more insidious role. It serves to distract from the very real problems of access and segretation to do exist in some parts of the city in particular concentrations - particularly in Whitehawk, but also in other parts of the city. Casting it as a problem for the whole city, serves the purpose of diverting attention away from the uncomfortable truth this is really a concentrated problem that needs a specific remedy for a particular part of the city - and not the one that’s on offer. And that in large parts of the city, disadvantaged children are doing better than in 2/3 of the rest of the country.
I know London is an entirely different context in many ways, but we should be striving to achieve something closer to the 42.1% FSM attainment rates achieved in Inner London - West and ignoring where the problems are most acute is not going to be the way to do it. And it absolutely should be in the Council’s mission to improve further, but in doing so, getting a better understanding of the various factors at play in London (demographic, financial, geographical, educational etc.) and the specific routes to success and turnaround is crucial, as then better solutions could be trialled here.
The Importance of Boundaries and distributions of people within them
I would like to thank Bea Taylor for helping me and producing the illustrated account below of the problems of statistical smoothing and the modifiable areal unit problem.
Boundaries - where they are drawn, who’s in, who’s out, are right at the centre of this issue. They are also important as they can be used to make statistics appear better or worse in a given situation.
One additional fear, which I didn’t articulate at the top of this piece in my fears section, is that if the Council propose Option B (or something that looks quite like Option B) one of the justifications they will try to use is “it’s the only option which sees an ‘improvement’ - i.e. more even distribution - in Free School Meals across the catchments.”
However, much of any ‘improvement’ in a situation like this, will be down to statistical smoothing, and in some cases, as the illustration shows below, the priority allocation of FSM places where there is choice in the system, will make things worse for less popular schools - something we are likely to see in Brighton.
Statistical smoothing
Something to watch out for in the options being presented by the council is statistical smoothing. This is what happens when you reduce the number of zones in the city (known as the Modifiable Areal Unit Problem (MAUP)), ‘zones’ here being school catchment areas.
Here’s a simple example to explain. Suppose we have four schools with four catchment areas. Let dots represent pupils, and for simplicity let’s assume that the school caters to all, and only, the pupils in the catchment.
Now suppose the catchments are re-drawn so each catchment area serves two schools, and schools randomly select pupils from the catchment according to school capacity.
Under the original catchments the % of pupils eligible for free school meals varied from 15%-60% across the schools. Under the new catchments the % of pupils eligible for free school meals varies from 30%-40% across the schools.
The number of pupils eligible for free school meals has not changed, and the distribution of where these pupils reside in the city has not changed.
The gap in the percentage distribution of pupils eligible for free school meals has shrunk, because all schools have moved closer to the city average for the % of pupils eligible for free school meals.
Something to bear in mind is that these new percentages depend on pupils being chosen completely randomly from the relevant catchment. In reality, rules come into play, for example accommodating siblings at the same school, which mean such a total equality of mixing doesn’t occur.
Free School Meals and Excess School Capacity
The council are, laudably, trying to improve the opportunities for pupils eligible for free school meals by distributing the pupils more evenly through the city. One way they’ve been trying to do this is by introducing a new rule for how schools accept pupils, whereby pupils eligible for free school meals get first dibs on their school of choice. However, due to the way excess capacity is distributed across the schools in Brighton, it will potentially only lead to already less popular schools becoming less popular still.
The best way to explain this is with another simple example. Let’s suppose we have a city with three schools and three catchments. Each dot on the map represents a pupil entering secondary school. The schools have varying capacity (equivalent to the PAN), and varying levels of popularity (denoted by stars) which could be thought of as reflecting school outcomes, or OFSTED rating, or geographic location.
School A is in the suburbs of the city. School B and C are in the centre of the city, in more densely populated areas. School A occupies a large site, with a large capacity for pupils, perhaps it was built in a period of high demand due to lots of families moving to the immediate area.
Now suppose this area around school A is now less densely populated by families, for example there might be an ageing population in this area. Simultaneously due to inequalities in the city, the distribution of pupils eligible for free school meals in not even across the city, for example perhaps due to rising rent prices in the city centre, lower income households have had to move to the suburbs – contributing to higher rates of pupils eligible for free school meals in this area.
Now lets think about how these pupils might be assigned to secondary schools in the city. Prior to assignment we have:
Now lets assume the school admits pupils according to the following rules, until the school’s maximum capacity is reached:
Rule 1: the school admits anyone who has chosen the school AND gets a free school meal.
Rule 2: The school admits anyone who has chosen the school AND lives in the catchment area.
Rule 3: The school admits anyone else who has chosen the school.
I’m going to make one final assumption that for ~90% of pupils their first choice is the school they are in the catchment for.
So according to Rule 1, we start by looking at where pupils eligible for free school meals go to school. Let’s assume that all pupils would normally choose the school they are in the catchment area for. However, under the new rule, which allows pupils eligible for free school meal to attend any school in the city, 2 pupils in the catchment for school A elect to attend school B instead.
After rule 1, School A has an intake of 4, School B an intake of 4, and School C an intake of 3.
Let’s assume all other pupils want to attend the school in their catchment area. School A has 16 places remaining, and admits the 6 remaining pupils in its catchment who chose it -resulting in 10 pupils total.
School B has 6 spaces remaining, it randomly admits 6/8 pupils from it’s local area before reaching capacity. This leaves 2 pupils whose school is yet to be determined (highlighted in red below).
School B has 11 places remaining, and admits the 9 remaining pupils in its catchment – resulting in 12 pupils total.
Returning to the two pupils whose school in undecided - School B is completely full and can no longer admit students, however School A and school C still have capacity. Since school C is almost as good as B, as well as being nearby in the city centre, these pupils elect school C – resulting in 14 pupils at school C.
Now looking at the table from earlier again:
In this example situation, both schools B and C have reached capacity. School A and B have a more equal distribution of pupils eligible for free school meals, but the percent of pupils eligible for free school meals has decreased at school C, since it has had an influx of pupils not on free school meals.
Now, in terms of excess capacity, this all lies in school A, it has filled just 50% of its potential capacity. When given a choice, pupils move to the better schools, or failing that the next best school until they reach capacity.
This is a really simplified example. How the school allocates places is more complicated. Consider the rule for children eligible for free school meals – the actual rule states that ‘the number of places in this priority are limited so that the school has no more than the average number of free school meals children attending.’
Any ‘improvement’ such as this would be merely statistical averaging. In Geography this is called and you can see it in operation in the map below - click on the layers to see how changing the boundaries (with exactly the same schools underneath) alters the FSM percentages.
Figure 2 - Statistical Averaging for Options A, B and C - FSM across Brighton.
Observations
In the Figure above, I have taken the published school level FSM percentages from the DfE annual database - these are published at School level and thus slightly different from the figures published by Brighton Council (again, please just make more of this data publicly available!)
Turning the different layers on and off, it’s possible to see the effects of this statistical averaging across the catchments - the underlying data remains EXACTLY the same - these are percentages of students on FSM in each school, just averaged differently across the different catchment options.
In Option A - the BACA catchment has the highest concentrations of FSM at 42%, closely followed by Longhill at 40%
In Option B - as BACA has now merged with Patcham and Varndean, it’s average is brought closer to theirs and to the city average at 26%. Longhill also appears to have improved to 27%.
However, NOTHING HAS CHANGED. It’s the same underlying data for the same schools. The only thing that is different is some statistical averaging. Imagine if we kept expanding the catchments until there was just 1 - covering the whole city. The FSM for this catchment would of course be the city average!
So any claims from the council that having larger catchments somehow magically FSM distributions needs to be questioned. If the pupils and the schools stay the same, it’s merely statistical averaging.
Now, of course, I am fully aware that in having larger catchments and random allocation to schools within those catchments, then there’s a good chance that there will be a lot more mixing, but this will take years to propagate through and the council can’t yet predict how that will play out, so be cautions of justifications based on statistical averaging.
Brighton and Hove - Demographic Patterns
History and Futures
One question that keeps coming up relates to what the school population of Brighton is going to look like in the future.
While, as I showed in my last piece, it’s fairly certain from ONS projections that the city is going to lose quite a number of young people over the next decade - maybe a couple of thousand - what is not clear is how that will play out across the city. Will it be an even decline, or will some places suffer worse than others?
These don’t quite go down to the LSOA level, but are available at the local authority level and go a number of years into the future.
What is noticable is the relative distributions of the LSOA populations don’t change that much over time, so we can use the distributions in the past estimates, to apply to future projections to create a full neighbourhood-level picture of Brighton for a 20 year period.
One of my colleagues, Dr Andy Maclachlan, has created an interactive map where you can explore the period from 2011 until 2031 for the city. https://amaclachlan.shinyapps.io/Brighton_school_children_catchments/ - all data have been downloaded from the ONS via the NOMIS website.
An alternative view of change can be seen in the map below.
# Selecting just the columns geography and geometry from BrightonLSOABrightonLSOA_selected <- BrightonLSOA %>%select(lsoa21cd, geometry)# Performing the left joinBrightonLSOA_selected <- BrightonLSOA_selected %>%left_join(BTN_LSOA_Proj_wide, join_by("lsoa21cd"=="geography"))# Checking the resultqtm(BrightonLSOA_selected)
# Selecting the relevant columns (assuming the columns are named "year_2011" to "year_2031")year_cols <-grep("20[1-3][0-9]", colnames(BrightonLSOA_selected), value =TRUE)# Pivoting the dataset to long formatBrightonLSOA_long <- BrightonLSOA_selected %>%pivot_longer(cols =all_of(year_cols),names_to ="year",values_to ="value" )# Converting year column to numeric if necessaryBrightonLSOA_long$year <-as.numeric(gsub("year_", "", BrightonLSOA_long$year))# Creating the cleaned-up faceted ggplot mapggplot(BrightonLSOA_long) +geom_sf(aes(fill = value), color =NA) +# Remove polygon boundaries with color = NAfacet_wrap(~ year, ncol =4) +theme_minimal() +theme(axis.text =element_blank(), # Remove axis textaxis.title =element_blank(), # Remove axis titlespanel.grid =element_blank() # Remove grid lines ) +labs(title ="11 year olds in Brighton over time",fill ="Value")
# Create the yearly_sum dataframeyearly_sum <- BrightonLSOA_long %>%group_by(year) %>%summarise(sum_value =sum(value, na.rm =TRUE))# Load necessary library for rolling meanlibrary(zoo)
Attaching package: 'zoo'
The following objects are masked from 'package:data.table':
yearmon, yearqtr
The following objects are masked from 'package:base':
as.Date, as.Date.numeric
# Calculate the 3-year rolling averageyearly_sum <- yearly_sum %>%mutate(rolling_avg =rollmean(sum_value, k =3, fill =NA, align ="right"))# Create the line graph with the rolling averageggplot(yearly_sum, aes(x = year)) +geom_line(aes(y = sum_value), color ="blue") +geom_line(aes(y = rolling_avg), color ="red", linetype ="dashed") +theme_minimal() +labs(title ="Count and 3-Year Rolling Average - 11 year olds - Brighton",x ="Year",y ="Count 11 year olds",color ="Legend") +scale_color_manual(values =c("Yearly Sum"="blue", "3-Year Rolling Average"="red"))
Warning: No shared levels found between `names(values)` of the manual scale and the
data's colour values.
Warning: Removed 2 rows containing missing values or values outside the scale range
(`geom_line()`).
One thing to point out is that while numbers in a given neighbourhood may fluctuate a little, areas don’t change very much relative to each other.
Between 2011 and up to 2031, we can see that the number of 11 year-olds in the city hit a peak in 2019 with another small peak in 2024, but there is a projected decline of 200-300 students
Demographic trends are long term generally, so the patterns for schools which draw from large parts of the city, can be fairly predictable. In the East of the City, for example, the interactive map shows that in the ‘Deans’ most LSOAs (except for one between Ovingdean and Woodingdean, which might be around 5 11 year-olds higher in 2031 than in 2011) there is nothing to suggest anything other than the same decline experienced elsewhere in the city.
Demographic Distribution
The crucial thing in Brighton, however, relates to the spatial distribution of school age Children.
I posted this map a few days ago as it is fundamental to understanding the problems in the City
# Create the tmaptmap_mode("view")
tmap mode set to interactive viewing
tm_shape(BrightonLSOA_pw_cent) +tm_dots(size ="2024", col ="2024", alpha =0.5, border.alpha =0, title ="Total Children") +tm_layout(legend.show =TRUE, title ="11 year olds, 2024") +tm_shape(bh_catchments) +tm_borders() +tm_fill(alpha =0) +tm_shape(brighton_sec_schools_sml) +tm_dots()
Warning: The shape bh_catchments is invalid (after reprojection). See
sf::st_is_valid
Legend for symbol sizes not available in view mode.
Figure - - The spatial distribution of 11 year olds in 2024, relative to current secondary school catchment boundaries.
In Figure x above, the circles relate to LSOA neighbourhoods in the city. The numbers relate to the 2024 sub-national population projections at City Level, distributed down to LSOAs using the last know recorded distributions in 2021. There is bound to be some error, but as you will see from the historical 2011 to 2021 patterns shown above and on the interactive site, demographics are stable enough to be confident this is a pretty fair representation of the real distribution of 11 year olds.
Normally these sorts of data are shown on a choropleth map of contiguous zones, but here I have chosen to take the population-weighted centroids of the LSOA zones and scale those centroid locations according to estimated numbers of 11-year olds. In doing this, the distorting effects of open space are controlled for and a truer picture of the real population distribution in the city can be gained.
Overlaid onto this map are the current secondary school catchment boundaries in the city.
It becomes very clear that most catchments are well populated although the BACA catchment in the north-east of the city appears to have a low population within its boundary. This is born out in the analysis I conducted in my last piece.
The other thing to note is how sparse the Longhill Catchment to the East is, with it clear that other than a small concentration of children to the north in Woodingdean, the majority of the catchment population is made up of children coming from Whitehawk to the West. Without those children from Whitehawk, it would be difficult to see Longhill having enough children to keep running in its current form.
